一种改进的增量数据挖掘算法

摘要

For the issue that present incremental data mining algorithms can't solve the increment updating problem of which the database alters as the minimum support and minimum confidence are changing, we analysed the insufficiency of existing algorithms,and put forward an improved algorithm, which makes fullplay of the previous mining results, combines chain insertion, deletion with the characteristic of high-efficiency of hash positioning,gets rid of the deficiency of multiple iteration the traditional algorithms have,and is able to achieve increment updating that when the minimum support and minimum confidence are changing,the database will also change simultaneously. The algorithm is validated through experiments,and its test result is compared with the test result of Apriori algorithm to have proved its high efficiency.%针对现行增量挖掘算法都不能解决最小支持度和最小置信度发生变化的同时数据库也发生变化的增量更新问题,分析现有算法的不足,提出一种改进算法.该算法充分利用以前挖掘的结果,将链表插入、删除以及hash定位高效率的特性结合起来,摆脱了传统算法多次迭代的不足,可实现最小支持度与最小置信度发生变化的同时,数据库也相应的发生变化的增量更新.并通过实验对该算法进行验证,同时将测试结果与Aprior算法的测试结果进行比较,证明其高效性.